Home - joectchan/blog GitHub Wiki

Embedded Systems

Zephyr RTOS

Time Series Data Science Class at Santa Clara Adult Education

Attended a class taught by Dan Bikle four years ago. website

2 day workshop - Deep Learning for NLP with PyTorch

Ubuntu PC

I have one monitor attached to my Windows laptop. My development are all done on a NUC x86 running Ubuntu. I keep it headless and use RDP from Windows to access Gnome desktop. I also setup VNC access to this Gnome desktop, so that I can choose either RDP or VNC.

This link talks about a lot of customizations I like. I add a few more for my own environment.

grub

I need to enable grub menu on boot because I have Ubuntu 19.10 and 20.04 on my NUC

Use sudo vi /etc/default/grub to add non-zero timeout.

GRUB_TIMEOUT_STYLE=menu
GRUB_TIMEOUT=30

Use sudo update-grub after finished editing.

ssh

Configure ssh server link

Do NOT follow their optional step of allowing ssh login as root!!

clamav and rkhunter

Install with sudo apt install clamav rkhunter

Use sudo vi /etc/default/rkhunter to change these.

CRON_DAILY_RUN="true"
CRON_DB_UPDATE="true"
REPORT_EMAIL="your username on your ubuntu"
APT_AUTOGEN="true"
NICE="19"

Use sudo vi /etc/rkhunter.conf to change these.

UPDATE_MIRRORS=1
MIRRORS_MODE=0
MAIL-ON-WARNING=your username on your ubuntu
WEB_CMD="curl"

Use sudo vi /etc/rkhunter.conf.local to add these for specific config for ubuntu and apps that I use, because rkhunter warns against them. I prefer to keep overrides in a separate file, instead of modifying /etc/rkhunter.conf.

PKGMGR=DPKG
ALLOWDEVFILE="/dev/shm/pulse-shm-*"
ALLOWHIDDENDIR=/etc/.java
ALLOWIPCPROC=/usr/bin/gnome-software
ALLOWIPCPROC=/usr/bin/nautilus-desktop
ALLOWIPCPROC=/usr/bin/nautilus
ALLOWIPCPROC=/usr/bin/evince
ALLOWIPCPROC=/usr/lib/firefox/firefox

This is how I started my first scan with rkhunter

sudo rkhunter --update
sudo rkhunter --propupd
sudo rkhunter --versioncheck

sudo rkhunter --check -sk

Besides the warnings that I ended up supressing with rkhunter.conf.local, it gave me a real warning about my poor config of ssd. I need to fix by

/etc/ssh/sshd_config

PermitRootLogin=no

I never need to config clamav. I just run it by clamscan -r -i

I setup cron job to run clamav and rkhunter daily. sudo crontab -e

0 0 * * * /usr/bin/rkhunter -c 2>&1 | mail -s "Report of My Laptop: rkhunter" <your email>
3 0 * * * /usr/bin/clamscan -r -i /root 2>&1 | mail -s "Report of My Laptop: clamscan /root" <your email>
6 0 * * * /usr/bin/clamscan -r -i /home 2>&1 | mail -s "Report of My Laptop: clamscan /home" <your email>

I know rkhunter scan is pretty fast. I don't use root account for anything so clamscan /root should be fast. I estimate they take 3 minutes each. Scan /home takes a long time...

  • Config clam This is a much more elaborated configuration of clamav

arandr

I have two monitors at work. I use arandr to config their rotation and placement setting. This is very useful when I am not setting up a headless Ubuntu.

sudo apt install arandr
arandr

xrdp

I setup my headless NUC for RDP to use the built-in rdp client on Windows 10.

  • Wifi connection drop will not hurt. Just reconnect.
  • RDP desktop size depends on RDP client size when making connection. VNC desktop size does not change.

Make my Ubuntu boots in text mode. Otherwise, my Gnome desktop will be display on the non-existing monitor and xrdp will not be able to show me a desktop remotely.

systemctl set-default multi-user.target

Reboot to confirm Ubuntu boots into text mode. Then install and confirm xrdp runs.

I will visit https://c-nergy.be and get their xrdp-installer script. I got mine under their Product tab, version 1.2 for ubuntu 20.04. My commands only installs xrdp and check it is running.

(Obsolete commands. Use the script I mentioned instead.)
sudo apt install xrdp
sudo systemctl status xrdp

ip addr show shows ip address of my NUC. Over on my Windows laptop, I use the built-in rdp client and enter the IP address of my NUC.

hostnamectl status shows the hostname of my NUC. (which is nuc in my case.) I can do a ping nuc.local on my Windows laptop to find the ip address of my NUC, thanks to avahi-daemon. link

VNC

VNC server is built-in to Ubuntu GUI desktop. If you keep the default graphical.target with systemctl, you can open Settings..Sharing. Enable sharing with the sliding button. Click Screen Sharing. Enable screen share with another sliding button.

Check this setup with command avahi-browse -a at the shell. You will see.

+   wlo1 IPv6 joseph's remote desktop on nuc                VNC Remote Access    local
+   wlo1 IPv4 joseph's remote desktop on nuc                VNC Remote Access    local

When I use another Ubuntu in my LAN to access my NUC machine, I can open Remmina and type nuc.local into the address bar of Remmina and press Enter. The avahi-browse command has already confirmed VNC server is advertising its service in my LAN. Hence, Remmina on another Ubuntu will be able to see that service in the same LAN.

The quality of graphics is poor by default because the VNC connection is using 8 bit color. I will quit and properly add a connection to my NUC. Edit color depth to High color and quality to Medium.

Remmina also support SFTP and SSH protocols. I like to add an sftp connection to NUC to facilitate file transfer between linux machines on my LAN.

Gnome Extensions and Gnome Tweaks

Open Terminal and type the following.

sudo apt install gnome-shell-extensions gnome-tweaks

Logout and login. Open Terminal and type the following.

gnome-tweaks

Under General, I turn off animation because I use xrdp to remote access. I need to cut down eye candy. Under Extensions, I recommend to use Workspace Indicator. Alternatively, I also like an extension called Workspaces to dock. link

UFW

Allow ssh, rdp and vnc connection within my LAN 192.168.1.X.

sudo ufw allow proto tcp from 192.168.1.0/24 to any port 22,3389,5900
sudo ufw enable
sudo ufw status numbered

Raspberry pi wifi scan

I installed KDE on ubuntu on a rpi. It bugs me for root password to do wifi scan all the time. That is because NetworkManager permission is not configured properly. sudo nmcli general permissions

PERMISSION                                                        VALUE
org.freedesktop.NetworkManager.wifi.scan                          unknown

This command shows the detail of this polkit policy sudo pkaction --action-id org.freedesktop.NetworkManager.wifi.scan --verbose

org.freedesktop.NetworkManager.wifi.scan:
  description:       Allow control of Wi-Fi scans
  message:           System policy prevents Wi-Fi scans
  vendor:            NetworkManager
  vendor_url:        http://www.gnome.org/projects/NetworkManager
  icon:              nm-icon
  implicit any:      auth_admin
  implicit inactive: yes
  implicit active:   yes

I create a new file 46-allow.wifi.scan.pkla in /etc/polkit-1/localauthority/50-local.d/ to get rid of the permission issue.

[Allow WiFi scan all Users]
Identity=unix-user:*
Action=org.freedesktop.NetworkManager.wifi.scan
ResultAny=yes
ResultInactive=yes
ResultActive=yes

Gnome shell extensions

They are installed by default in ~/.local/share/gnome-shell/extensions/

ls ~/.local/share/gnome-shell/extensions/

[email protected]

Enable an extension via command line by

gnome-shell-extension-tool -e [email protected]

This is useful to know for shell scripting some installation.

OpenGPG

I come across these pages showing to setup OpenPGP keys on windows. I will try to do that on Ubuntu instead in the future. Generate OpenPGP keys SSH with OpenPGP and YubiKey

IPv6

When ping nuc.local does not respond, i.e. mDNS somehow does not work, I use IPv6 instead to connect to my headless linux boxes. First, find the name of the ethernet interface of my local linux box.

ip addr show | grep "state UP"
3: wlp2s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UP group default qlen 1000

Ssh back to localhost via IPv6

ssh -6 ::1

Show the IPv6 addresses of this interface.

ip -6 addr show dev wlp2s0

or

route -A inet6 | grep  wlp9s0

Detect IPv6 neighbors

ip -6 neigh show

Alternatively, detect IPv6-active hosts on a link by using ping6 to the link-local all-node multicast address.

ping6 -I wlp2s0 ff02::1

Ssh to the neighbor, the headless linux box, using its IPv6 address and the local interface used to connect to that address.

ssh -6 fe80::fc42:3cdd:7c2c:57c%wlp2s0

Follow this link to configure Avahi with IPv6

Turn off laptop screen in text console

Since I use my old laptop as headless linux box, I want to turn off the LCD screen. Configure using DPMS in text conssole. Login at my laptop into a console. Enter this command to force the LCD screen off.

setterm -blank force -powerdown 1 -powersave powerdown

Serial port permission

Link shows the steps to grant myself access to serial port. It is basically adding myself to the dialout group and reboot.

VPN using sshuttle

Link

Tech links

Periodic

Tech Articles

git

Data structure and Algorithm

C links

Debug

  • GNU poke Analyze binary file
  • linkSteps to separate debug symbol out of elf binary
  • medium linkCreating and using debug symbol tables with CMake and GDB
  • pretty print STL in gdb
  • Recipe of config pretty print STL in gdb on ubuntu
  • Add python modules to the python interpreter inside gdb
  • Config VSCode debug with gdb and pretty print STL
  • [Summary of Debugging Tools for Parallel Applications] (http://pramodkumbhar.com/2018/06/summary-of-debugging-tools/)
  • Config ddd to use modern gdb and custom pretty printer ddd can record history of a watched variable. Use python script to test variable type and custom pretty print your datatype.
  • Gotcha is a library that wraps functions. Tools can use gotcha to install hooks into other libraries, for example putting a wrapper function around libc's malloc.
  • GDB dashboard is a standalone .gdbinit file written using the Python API that enables a modular interface showing relevant information about the program being debugged.
  • Write up about gdbgui. It can "monitor the value of certain variables over time on a chart." This is the installation guide
  • link ARM Micro Trace Buffer and gdb python script to interpret buffer data.
  • link Debug ARM hard fault
  • link Describe fault handler and PyCortexMDebug gdb py script to parse svd file and apply in gdb
  • pdf Tracing Summit 2018 Evolution of self-hosted Tracing with ARM CoreSight
  • pdf Hardware Assisted Tracing on ARM with CoreSight and OpenCSD, by Mathieu Poirier
  • link Hardware Tracing for Fast and Precise Performance Analysis. Intel and ARM
  • link Intel CPU tracing tool vendor books.
  • link One blog post from the above vendor. Using the Intel Trace Hub for at-speed printf
  • link 2019 Explain Intel Processor Trace
  • Linaro OpenCSD lib Decoding CoreSight trace.
  • pdf 2014 Hardware-Assisted Software Event Tracing, performance
  • link MIPI System Software - Trace (MIPI SyS-TSM) is a common data format for transmitting software trace and debug information between a test system and a device, such as a system-on-chip (SoC) or platform.
  • link 2016 Linaro blog. System Trace Module (STM) and its usage. Has link to a sample dtsi file for Spreadtrum’s SC9836
  • linkLinux 5.4.47 Documentation/devicetree/bindings/arm/coresight.txt
  • link Trace capture/decode with CoreSight driver and Perf on Linux, using arch/arm64/boot/dts/arm/juno-base.dtsi board
  • Generate backtrace with abseil-cpp
  • Print stack trace example
  • Print stack trace example
  • Print stack trace example
  • backward-cpp C++ header only library to print stacktrace
  • Icecream-cpp header only debugging lib for cpp
  • Embedded C trace Trace macro. Tool to strip trace text and only transmit binary. Pair with json to reconstruct the message on host.
  • PerfettoThe Perfetto Tracing SDK is a C++11 library that allows userspace applications to emit trace events and add more app-specific context to a Perfetto trace.
  • Eclipse Trace Compass Visualizing CommonTraceFormat, gdb trace
  • uftrace Capture function trace of a user space application compiled with -pg. Slide 5 shows a list of libraries to compile uftrace.
  • Linux Full Stack Tracing Cross-compile ARM linux kernel and install Lttng tools. Use shell scripts to start and stop Lttng capture.
  • metal-ci Bare metal uniting testing. Python-gdb script helps to print the values on the target when test failed. Or just serial port unit testing, where serial port message shows a test condition failed without providing the runtime value(s) failing that test.

Performance

Buses links

USB

I3C

IO-Link

CAN

ARM

C/C++ programming

ZBar is an open source software suite for reading bar codes from various sources, such as video streams, image files and raw intensity sensors. It supports many popular symbologies (types of bar codes) including EAN-13/UPC-A, UPC-E, EAN-8, Code 128, Code 39, Interleaved 2 of 5 and QR Code.

The flexible, layered implementation facilitates bar code scanning and decoding for any application: use it stand-alone with the included GUI and command line programs, easily integrate a bar code scanning widget into your Qt, GTK+ or PyGTK GUI application, leverage one of the script or programming interfaces (Python, Perl, C++) ...all the way down to a streamlined C library suitable for embedded use.
C++ code 
    // Producer thread:
    Data_t g_data1, g_data2, g_data3;
    std::atomic<bool> g_flag

    g_data1 = calc1();
    g_data2 = calc2();
    g_data3 = calc3();
    g_flag.store(true, std::memory_order_release);

    // Consumer thread:
    if (g_flag.load(std::memory_order_acquire)) {
      DoSomething(g_data1, g_data1, g_data2, g_data3);
  • Protobuf enum as string
  • My own c++ example code showing the result of troubleshooting class member definition of a template class and using std::move() to pass functor into std::thread().
C++ code

Execute online with Coliru

#include <iostream>
#include <utility>
#include <thread>
#include <chrono>
#include <memory>
#include <syncstream>

// TODO: Reference additional headers your program requires here.

/*
Example 1: If a functor has data member of the type std::unique_ptr, then
            passing functor object to std::thread() REQUIRES
            either      std::move(functor object) in order to invoke the default move constructor
            or          define your copy constructor explicitly
*/
struct Vec3
{
    int x, y, z;
 
    // following constructor is no longer needed since C++20
    Vec3(int x = 0, int y = 0, int z = 0) noexcept : x(x), y(y), z(z) { }
 
    friend std::ostream& operator<<(std::ostream& os, const Vec3& v) {
        return os << "{ x=" << v.x << ", y=" << v.y << ", z=" << v.z << " }";
    }
};

class foo
{
public:
    foo() { std::cout << "foo()" << " count: " << ++count << " n: " << n << '\n'; };
#if 1
    // This functor class does not have a unique_ptr as a data member.
    // We can pass it into thread by using:
    //      std::thread(foo)
    // without writing our custom copy constructor. Default copy constructor works.
    // You can change the #if to zero to try that.
    foo(const foo& org) : n(org.n) { std::cout << "foo(const foo&)" << " count: " << ++count << " n: " << n << '\n'; };
    foo(foo&& org) : n(std::exchange(org.n, 0)) { std::cout << "foo(foo&&)" << " count: " << count << " n: " << n << '\n'; };
#else
    foo(const foo&) = default;
    foo(foo&&) = default;
#endif    
    
    void operator()()
    {
        std::osyncstream synced_out(std::cout);
        v1_ = Vec3{3,4,5};
        for (int i = 0; i < 5; ++i) {
            synced_out << "Thread foo executing\n";
            ++n;
            std::this_thread::sleep_for(std::chrono::milliseconds(10));
        }
        synced_out << "Vec3{3,4,5}: " << v1_ << " n: " << n << '\n';        
    }
    
    int n = 10;
    static int count;
    
private:
    Vec3 v1_;
    
};


 
class baz
{
public:
    baz() {std::cout << "baz()" << " count: " << ++count << " n: " << n << '\n'; };
#if 1
    // This functor class has a unique_ptr as a data member.
    // We can pass it into thread by:
    //      std::thread(std::move(baz))
    // Default move constructor works.
    //
    // Or write and a custom copy constructor that copy everything except the unique_ptr.
    // Then we can continue to use the syntax
    //      std::thread(baz)
    baz(const baz& org) : n(org.n) { std::cout << "baz(const baz&)" << " count: " << ++count << " n: " << n << '\n'; };
    baz(baz&& org) : n(std::exchange(org.n, 0)), v2_(std::move(org.v2_)) { std::cout << "baz(baz&&)" << " count: " << count << " n: " << n << '\n'; };
#else
    baz(const baz&) = default;
    baz(baz&&) = default;
#endif    
    
    void operator()()
    {
        std::osyncstream synced_out(std::cout);
        v2_ = std::make_unique<Vec3>(0,1,2);
        for (int i = 0; i < 5; ++i) {
            synced_out << "Thread baz executing\n";
            ++n;
            std::this_thread::sleep_for(std::chrono::milliseconds(10));
        }
        synced_out << "make_unique<Vec3>(0,1,2): " << *v2_ << " n: " << n << '\n';        
    }
    
    int n = 20;
    static int count;
    
private:
    std::unique_ptr<Vec3> v2_ = nullptr;
    
};

int foo::count = 0;
int baz::count = 0;

/*
Please read main() to understand the point made by example 1. Try uncommenting thread 4 code.
*/


/*
Example 2: Class data member of a template class can only be defined with an initializer.
            Without an initializer, that statement is only a declaration statement.
*/

namespace grpc {
class Channel {
public:
    Channel(const std::string& target) { std::clog << "Channel to " + target + " | class data member, implicit_init_ptr: " << implicit_init_ptr << '\n'; };
private:
    static int* implicit_init_ptr;
};

int global_integer = 42;

#if 1
// Non-template class can define class data member by implicit initialization
int* Channel::implicit_init_ptr;
#else
// Non-template class can also define class data member with explicit initialization
int* Channel::implicit_init_ptr = &global_integer;
#endif

std::shared_ptr<Channel> CreateChannel(const std::string& target) {
    return std::make_shared<Channel>(target);
};

} // namespace grpc

namespace ws {
    class Op {
    public:
        Op() { std::clog << "Op()\n"; };
    };

    class Api {
    public:
        Api() { std::clog << "Api()\n"; };
    };
} // namespace ws

template <typename T>
class WebServiceGrpcClient {
public:
    static void ini_channel(const std::string& host) {
		WebServiceGrpcClient<T>::channel_ =
			::grpc::CreateChannel(host);
    };

private:
    // All stubs (threads) share a single channel to hwmond.
    static std::shared_ptr<::grpc::Channel> channel_;
};

#if 0
// Template class can NOT define class data member by implicit initialization. These lines will be treated as declarations. Compilation will fail.
//
// main.cpp:(.text.startup+0x5b): undefined reference to `WebServiceGrpcClient<ws::Op>::channel_'
// main.cpp:(.text.startup+0x159): undefined reference to `WebServiceGrpcClient<ws::Api>::channel_'
//
template <>
std::shared_ptr<::grpc::Channel> WebServiceGrpcClient<::ws::Op>::channel_;
template <>
std::shared_ptr<::grpc::Channel> WebServiceGrpcClient<::ws::Api>::channel_;
#else
// Template class MUST define class data member with explicit initialization. These lines can compile and run
template <>
std::shared_ptr<::grpc::Channel> WebServiceGrpcClient<::ws::Op>::channel_ = nullptr;
template <>
std::shared_ptr<::grpc::Channel> WebServiceGrpcClient<::ws::Api>::channel_ = nullptr;

#endif

/*
End of example 2
*/

int main()
{
	WebServiceGrpcClient<::ws::Op>::ini_channel("op.web.service.com");
	WebServiceGrpcClient<::ws::Api>::ini_channel("api.web.service.api");

    int n = 0;
    foo f;
    baz b1;
    baz b2;
    
    f.n = 100;
    b1.n = 30;
    b2.n = 40;
    
    std::thread t1(std::move(f)); //  Default move constructor does not destory f.n. Our custom move constructor use std::exchange() and destory f.n after move.
    std::thread t2(f); //  Use f after std::move() is bad! Doing this just to observe the behavioral difference between the default and our custom  move constructor
    std::thread t3(std::move(b1)); //  //  Default move constructor does not destory b1.n. Our custom move constructor use std::exchange() and destory b1.n after move.
    std::thread t4(b2); // Compilation error when using the default copy constructor. Unique pointer has no copy constructor. Our custom copy constructor deliberately skips the unique pointer and that make this line work again.
    t1.join();
    t2.join();
    t3.join();
    t4.join();
    
    std::cout << "Final value of n is " << n << '\n';
    std::cout << "Final value of f.n (foo::n) is " << f.n << '\n';      //  Use after std::move() is bad! Doing this for educational purpose.
    std::cout << "Final value of b1.n (baz::n) is " << b1.n << '\n';      //  Use after std::move() is bad! Doing this for educational purpose.
    std::cout << "Final value of b2.n (baz::n) is " << b2.n << '\n';

  return 0;
}

Online C++ compiler

C++ IDE

CMake

Container

Golang links

Rust links

Python links

$python3 -m pip install --user pipx
Collecting pipx
  Using cached https://files.pythonhosted.org/packages/f6/7c/db71a08fe11bf509bd9fa3b6d89066dc5e8d0623f9280c7c235b4e012103/pipx-0.16.1.0-py3-none-any.whl
Will not install to the user site because it will lack sys.path precedence to colorama in /opt/python3.6c/lib/python3.6/site-packages

$ python3 -m pip install --user --no-deps pipx 
Collecting pipx
  Using cached https://files.pythonhosted.org/packages/f6/7c/db71a08fe11bf509bd9fa3b6d89066dc5e8d0623f9280c7c235b4e012103/pipx-0.16.1.0-py3-none-any.whl
Installing collected packages: pipx
Successfully installed pipx-0.16.1.0

pip install pyyaml

import yaml
try:
    with open (path_to_yaml, 'r') as file:
    config = yaml.safe_load(file)
except Exception as e:
    print('Error reading the config file')

python snippets

One line functions for pdb to save object as a text file using pretty print

def py_pp(o, fname, hdg): from pprint import pprint; outfile = open(fname+'.txt', 'w'); pprint(hdg, stream=outfile); pprint(o, stream=outfile); outfile.close()

def py_pp_vars(o, fname, hdg): from pprint import pprint; outfile = open(fname+'.txt', 'w'); pprint(hdg, stream=outfile); pprint(vars(o), stream=outfile); outfile.close()

pdb can accept multiline command. Use interact to enter a shell context. You can continue to access the variables you want to examine. In this example, I can print a desc_list, which contains objects of type fields. I print the name of the field object and then the content of the field object.

interact  
from pprint import pprint  
for rd in desc_list:
  print ('---')	
  for f in rd.fields:
    print (f.name)
    print ('---')	
    pprint (vars(f))

bash shell snippets

run command 10 times

for i in {1..10}; do commandNameHere; done

Mem and cpu usage of node

ps -C node -o pid,rss,pcpu,comm

Mem usage of users on the system

ps hax -o rss,user | awk '{a[$2]+=$1;}END{for(i in a)print i" "int(a[i]/1024+0.5);}' | sort -rnk2
  • link logical op and shell expansion
  • link Shell scripting. awk perl array whiptail
  • link File browser using bash and whiptail
  • link Pure bash script for user to build a menu for user to select
  • link Bash cut part of a string
  • link Command grouping with {cmd1; cmd2 ; cmd3} > single_output.txt or launch subshell with (cmd1; cmd2 ; cmd3) > single_output.txt

gRPC links

Linux links

  • rsyslog Setup remote syslog
  • rsyslog More examples setting rsyslog
  • rsyslog More examples setting rsyslog. Syntax seems to be newer.
  • Evolution Set up Evolution to work with MS Office365 account
  • LXD container manager
  • Linux homebrew homebrew for linux. Install package for single user without sudo. If your linux system does not have ruby pre-installed. Then use rvm to installl ruby for single user.If your system is old, do not brew update. I could get trufflerruby and linuxbrew running on an old pc with fedora 25. I attempted brew update then the newer brew code wants to access ruby in (/usr/bin:/bin:/usr/sbin:/sbin). That defeats my goal of single user install without sudo. At least I have tried. Oh well.
  • RVM This shows how to install rvm. Basically:
command curl -sSL https://rvm.io/pkuczynski.asc | gpg2 --import -
command curl -sSL https://rvm.io/mpapis.asc | gpg2 --import -
curl -sSL https://get.rvm.io | bash -s -- --version latest
source ~/.rvm/scripts/rvm
rvm list known

# My test pc has an old fedora. It does not have pre-built binaries for ruby 2
rvm install ruby-2.3
Searching for binary rubies, this might take some time.
No binary rubies available for: fedora/25/x86_64/ruby-2.3.8.
Continuing with compilation. Please read 'rvm help mount' to get more information on binary rubies.
Checking requirements for fedora.
Installing requirements for fedora.
Installing required packages: libyaml-devel..josephc password required for 'dnf install -y libyaml-devel': 

#I have to use truffleruby
rvm install truffleruby
Checking requirements for fedora.
Requirements installation successful.

Ran into this error when testing PyIDM, a tk and python app:

X Error of failed request:  BadLength (poly request too large or internal Xlib length error)
  Major opcode of failed request:  139 (RENDER)
  Minor opcode of failed request:  20 (RenderAddGlyphs)
  Serial number of failed request:  3026
  Current serial number in output stream:  3058

Ended up fixing the error by installing font to ubuntu 18.04 and 20.04

sudo apt install fonts-symbola

I had installed quite a number of prerequisites before PyIDM could run. This is my installation steps:

git clone https://github.com/pyIDM/PyIDM.git
sudo apt install ffmpeg fonts-symbola libcurl4-openssl-dev libssl-dev python3-pip python3-pil python3-pil.imagetk python3-tk 
python3 setup.py install --user
python3 -m pyidm

RTOS links

Micropython links

STM32 links

Espressif ESP32

AI links

Embedded software links

Hardware links

Android links

iOS links

Graphing links

  • ZenUML Online UML tool whose language syntax is similar to c++/java
  • Mermaid Online UML tool whose language syntax is similar to markdown
  • Simple intro to graphviz language
  • online graphviz 1 Good gallery
  • online graphviz 2 Export different output format
  • online graphviz 3 Zoom in/out preview. Undo/redo edit view.
  • link Text to UML tools – Fastest way to create your models
  • Gaphor Desktop UML graphical editor
  • link PlantUML with VSCode
  • link PlantUML Hitchhikers Guide
  • link Support UML diagrams in markdown with plantuml syntax
  • link Using plantuml in AsciiDoc on GitHub
  • Borwser extension PlantUML Visualizer extension
  • link Use PlantUML proxy service to render UML for github page
  • link How to integrate UML diagrams into GitLab or GitHub
  • link A Dramatic Tour through Python’s Data Visualization Landscape (including ggplot and Altair)
  • Altair
  • Plotly Dash

Windows links

WSL2 links

Security

Quantum computing

Programmer interview

Logging library

fmtlibUse fmtlib as header only library

C++ and other lanaguage library

Tech book

Tool to read csv

  • Visidata This is a ncurse type python program. It reads csv, json, ... etc and present a grid of data (a sheet). You can select rows based on some criteria and spawn a new sheet. You can define new columns using Python expression. You can join two sheets.
  • q - Text as Data Run SQL directly on csv files

Other links

Blog

Financial

  • Credit freeze; Opt out credit increase offer; Annual credit report Link

Car

Math

Books

Date / Time since 1AD, epoch, etc

Markdown

Restructured Text

Asciidoc

  • Its homepage Go to the "Try AsciiDoc on the Web" section of the homepage for online tools to try (AsciiDocBox and DocGist)
  • An opinion why use AsciiDoc

Wiki comparison

Build Github pages with

Build Github pages with Gatsby

Build Github pages with Jekyll

I spent an entire day trying out many themes. I ended up using Minimal Mistakes Other themes gave me too many problems. I had to delete my https://github.com/joectchan/joectchan.github.io repo to start from scratch.

Use this to clone mmistakes as my new https://github.com/joectchan/joectchan.github.io repo. Use this to create a favicon.ico and save it at the root of my repo because mmistakes does NOT provide a favicon.ico.

I ignored the entire GETTING STARTED section and plunged into CUSTOMIZATION.

Resources for building website with Jekyll

Online games

  • Jitsi WebRTC base video conference. Just use browser on computer. No download no account is needed. (Android and iOS needs to download mobile app, however.

  • Netgames Love letter, Avalon, Secret Hitler, ... Just use browser on mobile phone. No download no account is needed.

  • Maque 4 player mahjong. Can either play as guest or register for a free account.

  • Free or paid boardgames online

  • Paid Dominion cardgame online

  • List of online games

⚠️ **GitHub.com Fallback** ⚠️